how to ues t_tide

Our teacher asked us to calculate the amplitude and phase of tidal components with t_tide. I have downloaded the t_tide package, and how to use it, what are the inputs,，can it input multiple sets of data?please input any data [nameu,fu,tidecon,xout]=t_tide(xin,varargin)

i write,

Answer 1

[William Rose]

@ke, the t_tide package includes may functions including t_tide.m. The code for t_tide.m has extensive comments describing its inputs and outputs. For t_tide, the one required input is a time series - in other words, a vector or array of observed values at regular untervals (assumed to be 1 hour by default, but you can specify an alternative interval). The time series could be water elevations, water pressures, or current velocities (expressed as complex numbers), for example. There are various optional arguments, all described in the comments of t_tide.m.

See the article describing the t_tide package, here, for a more complete description. You should be able to access the full article via your university library.

Sounds like a good assignment. Good luck.

[ADRIEN BOUGUES]

Hello William, i'm searching where dowload the package t_tide, do you have a link ?

Thanks

[William Rose]

@ADRIEN BOUGUES,

https://www.eoas.ubc.ca/~rich/#T_Tide

Give credit to the authors Mike Foreman, Rich Pawlowicz, Steve Lentz, and Bob Beardsley.

[ADRIEN BOUGUES]

Thx, William

[Haikal Fayyadh Munadi]

Hello William, do t_tide can be use for altimeter data ? I'm trying to use t_tide for altimeter but I've been stuck on intervals issue for quite sometimes now. Could help me with that ?

[William Rose]

@Haikal Fayyadh Munadi,

You should be able to use t_tide with altimeter data. You said you have had issues with the sampling intervals. Can you post some data?

Have you been able to access the paper here? Your university should be able to assist with access.

See the web site of Rich Pawlowicz for updates on t_tide. The latest version is available here: https://www.eoas.ubc.ca/~rich/t_tide/t_tide_v1.5beta.zip.

[William Rose]

If your data is uneveny sampled, I recommend interpolating it to a uniform sampling rate, using interp1(). You can specify the interpolation method, and I recommend spline or makima. Method 'spline' is more smooth, in the sense of being limited to cubic polynomial in each interval. Method 'makima' avoids overshoots between points.

[Haikal Fayyadh Munadi]

echo on
echo on

% Define inference parameters.
infername=['P1';'K2'];
inferfrom=['K1';'S2'];
infamp=[.33093;.27215];
infphase=[-7.07;-22.40];

% The call (see t_demo code for details).
[tidestruc,pout]=t_tide(sla_untide,...
    'interval',237, ...                     % 9 day 21 hour
    'start',t,...               % start time is datestr(tuk_time(1))
    'latitude',6.272,...               % Latitude of obs
    'inference',infername,inferfrom,infamp,infphase,...
    'shallow','M10',...                   % Add a shallow-water constituent 
    'error','linear',...                   % coloured boostrap CI
    'synthesis',1);                       % Use SNR=1 for synthesis. 

echo off

[Haikal Fayyadh Munadi]

This is the code and sample of my data.

[William Rose]

@Haikal Fayyadh Munadi, the file you uploaded contains the scalar t=733600 and a vector of 280 points. Five of those points are NaN. Is t the sampling interval? In what units? If t is not the sampling interval, then what does t represent, and what is the sampling interval?

[Haikal Fayyadh Munadi]

yes. t is time for satellite altimeter data recorded.. I use datenum because I just follow the demo file. the sampling interval I put on is around 9 days but I change it into hours..

[William Rose]

@Haikal Fayyadh Munadi,

The code you posted includes

'interval',237, ... % 9 day 21 hour

The .mat file you posted includes the scalar t=733600. If that is the interval, in seconds, then the interval t equals approximately 203.7781 hours, or 8 days, 11 hours, 46 minutes, 41.2 seconds. Which is the correct interval: (9 days 21 hours) or (8d, 11h, 46m, 41.2s)?

Your data set has 280 values. If we assume the sampling interval is 9 days, then this data set corresponds to almost 7 years of data. Therefore it is essential that the sampling interval be known exactly, because even a small error will results in a large error in the phase of the lunar driving force, after 7 years. If the sampling intervals are even slightly uneven, then it will be important to provide a vector containing the date-times of all observations.

[William Rose]

@Haikal Fayyadh Munadi, please contact me if you woudl like to discuss this further. Contact me by clicking on my name above, then click the envelope icon in the pop-up window.